{ "architectures": [ "MiniCPMVForCausalLM" ], "auto_map": { "AutoConfig": "configuration_minicpmv.MiniCPMVConfig", "AutoModel": "modeling_minicpmv.MiniCPMVForCausalLM", "AutoModelForCausalLM": "modeling_minicpmv.MiniCPMVForCausalLM" }, "model_type": "minicpmv", "llm_config": { "model_type": "minicpm", "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 1, "dim_model_base": 256, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2304, "initializer_range": 0.1, "intermediate_size": 5760, "max_position_embeddings": 4096, "num_attention_heads": 36, "num_hidden_layers": 40, "num_key_value_heads": 36, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 10000.0, "scale_depth": 1.4, "scale_emb": 12, "scale_resolution": 448, "tie_word_embeddings": false, "use_cache": true, "vocab_size": 122753 }, "vpm_config": { "hidden_size": 1152, "image_size": 980, "intermediate_size": 4304, "model_type": "siglip_vision_model", "num_attention_heads": 16, "num_hidden_layers": 27, "patch_size": 14 }, "query_num": 64, "drop_vision_last_layer": true, "torch_dtype": "bfloat16", "slice_mode": true, "max_slice_nums": 9, "mm_use_im_start_end": true, "transformers_version": "4.40.0" }